vikingerik
Active member
- Nov 6, 2013
- 1,205
- 0
This started in the SPA Kickstarter thread but deserves a separate discussion.
Holy crap, I had no idea there's physics data human-readable in the game files. That's a gold mine of information, if we can figure out how to make use of it.
And also kinda disappointing, that TPA's physics "engine" is even more predestined and railroady than we already thought, if the possible trajectories from a flipper really are predefined that way.
I've also noticed that can't-miss feeling, and most particularly on Medieval Madness and CFTBL as we've mentioned. The big tell on CFTBL is that the same railroad that hits the left snackbar targets then flies up the ramp occurs exactly the same way coming from either a previous ramp shot or a snackbar eject. There's no way that should be true as emergent behavior, the ramp-rolling ball can't possibly behave exactly the same as the ejected one. Yet somewhere either in the feed or on the flipper itself, those converge to the same railroad.
This is very true. Aside from MM and CFTBL specifically, we were getting this sort of improvement at least as early as Twilight Zone. That shows quite a variety of behavior that feels emergent in trajectories up (and, importantly, nearly-up) its ramps and into the lock lane and camera.
Here's my hypothesis now. Somewhere in those parameters, perhaps elasticity, there's a weighting between the predefined trajectories versus including or ignoring the incoming ball velocity. It can be weighted either for all-predefinition like on MM and CFTBL, nearly so like on Funhouse and Monster Bash which also yield that "can't-miss" feeling, or weighted more towards emergent behavior like on TZ. Later tables that feel stiff and poorly tuned like HRC and Junkyard rely too much on predefinition. It can even be weighted per-flipper where Road Show's main flippers feel emergent but the blast-hole flipper feels easy and predestined, or even per-*shot* (per "transfer point" on the flipper) where Dracula's ramp feels unmissable while the mystery scoop needs an exact railroad from a catch. Want to see if you can corroborate any of that from the game data?
And people are conflating a higher weight on emergent behavior found in the later tables with "Physics 3.0". Or Farsight may even have intended that association.
This is taken/extracted from the TFnHouse.Rez file, PC/Steam revision 1.41.7, which was right after the change was made to make possible ~8 previously-impossible shots, and I think this is the change to Physics 3.0 but I'm not sure:
Code://1(length) 2(angle speed down) 3(elasticity when flipper static) 4 (elasticity when dropping) 5(bottom speed mult) 6(number of trnsfer pts) Physics, UpperLeft, 350, 26, 0.5, 0.25, 0.3, 10, TransferPoint2, 0.13, 24, 110, 200, 0, 0, 220, 250, TransferPoint2, 0.20, 26, 220, 270, 0, 0, 250, 320, TransferPoint2, 0.24, 28, 290, 270, 0, 0, 250, 320, TransferPoint2, 0.35, 32, 340, 270, 0, 0, 280, 320, TransferPoint2, 0.55, 36, 360, 320, 0, 0, 295, 360, TransferPoint2, 0.77, 42, 370, 350, 0, 0, 320, 390, TransferPoint2, 0.86, 45, 335, 320, 0, 0, 285, 360, TransferPoint2, 0.92, 50, 285, 270, 0, 0, 245, 310, TransferPoint2, 0.98, 60, 220, 200, 0, 0, 200, 250, TransferPoint2, 1.00, 70, 140, 150, 0, 0, 140, 200,
Holy crap, I had no idea there's physics data human-readable in the game files. That's a gold mine of information, if we can figure out how to make use of it.
And also kinda disappointing, that TPA's physics "engine" is even more predestined and railroady than we already thought, if the possible trajectories from a flipper really are predefined that way.
I've also noticed that can't-miss feeling, and most particularly on Medieval Madness and CFTBL as we've mentioned. The big tell on CFTBL is that the same railroad that hits the left snackbar targets then flies up the ramp occurs exactly the same way coming from either a previous ramp shot or a snackbar eject. There's no way that should be true as emergent behavior, the ramp-rolling ball can't possibly behave exactly the same as the ejected one. Yet somewhere either in the feed or on the flipper itself, those converge to the same railroad.
This is very true. Aside from MM and CFTBL specifically, we were getting this sort of improvement at least as early as Twilight Zone. That shows quite a variety of behavior that feels emergent in trajectories up (and, importantly, nearly-up) its ramps and into the lock lane and camera.
Here's my hypothesis now. Somewhere in those parameters, perhaps elasticity, there's a weighting between the predefined trajectories versus including or ignoring the incoming ball velocity. It can be weighted either for all-predefinition like on MM and CFTBL, nearly so like on Funhouse and Monster Bash which also yield that "can't-miss" feeling, or weighted more towards emergent behavior like on TZ. Later tables that feel stiff and poorly tuned like HRC and Junkyard rely too much on predefinition. It can even be weighted per-flipper where Road Show's main flippers feel emergent but the blast-hole flipper feels easy and predestined, or even per-*shot* (per "transfer point" on the flipper) where Dracula's ramp feels unmissable while the mystery scoop needs an exact railroad from a catch. Want to see if you can corroborate any of that from the game data?
And people are conflating a higher weight on emergent behavior found in the later tables with "Physics 3.0". Or Farsight may even have intended that association.
Last edited:
